Analog to digital converters (ADCs) are used to convert an analog input signal to an equivalent digital output signal. ADCs are required in a variety of applications where an analog signal must be converted to a digital signal in order to allow digital data processing.
Most digital data processing systems are not able to directly use an analog signal. Therefore, most analog values must be converted to an equivalent digital value by an ADC before a digital system can use the value. As a consequence, most digital data processing systems include an ADC if analog values are to be received by the system. Some microcomputer chips even include an ADC on board the microcomputer chip.
Fault grading is performed on a semiconductor chip design in order to increase reliability. The goal of fault grading is to generate tests which fully exercise all of the thousands or millions of individual devices and nodes within the chip in order to ensure that all of the devices and nodes function properly. Fault grading increases the reliability of chips which are manufactured using the fault graded design.
One set of nodes that is difficult to fault grade in a chip design is the first electrode of the capacitors used in an ADC. Fault grading of this set of nodes not only allows the capacitors themselves to be tested for various types of shorts, but in addition, allows the drive logic used to drive the first electrode of each capacitor to be tested. If resistors are used in place of some or all of the capacitors, then the set of nodes that is difficult to fault grade is the intermediate nodes between each resistive value. One of the reasons for the difficulty is that the voltages on these nodes are not digital voltages, but rather are analog voltages.
It is difficult to test and fault grade an analog structure such as the capacitors and resistors used in an ADC because the normal inputs to these analog structures are analog voltages. These analog input voltages cannot easily be generated with digital test equipment that uses only digital inputs and outputs. The digital testing equipment for a digital data processing system, such as a microcomputer chip, uses primarily digital input and output signals. The testing equipment may be able to use analog input and output signals, but such analog testing generally requires more testing time. Thus it is generally more efficient and less costly to perform tests that use digital rather than analog input and output signals.
In addition to fault grading the ADC design, each ADC chip that is manufactured must be individually tested using a linearity test, such as a full ramp test. This linearity test uses analog inputs to test the accuracy of the ADC over the full range of possible input and output values. A linearity test requires the ADC to perform hundreds or thousands of time-consuming analog to digital conversions. Linearity testing requires a great deal of time and is thus very costly. Therefore, it is desirable to find a way to test ADCs that requires less time, uses minimal additional circuitry in the ADC, and uses digital rather than analog inputs.
In present ADCs, the linearity test is often the only test method used to determine if the capacitors and/or resistors are defective. A linearity test requires the ADC to perform an analog to digital conversion for every inputted analog value. A partial linearity test is a linearity test which uses as imputs only a selected group of analog values within the input signal range. Although a partial linearity test is used to save test time in determining if the capacitors and/or resistors are defective, it still requires a significant number of analog to digital conversions.
A defect in the capacitors and/or resistors which determine the most significant bits of the digital output is easy to detect using a partial linearity test because an error in the most significant bits will not be masked by small errors due to other causes. However, a defect in the capacitors and/or resistors which determine the least significant bits of the digital output is harder to detect because an error in the least significant bits may be masked by small errors due to other causes, such as component mismatch. As a result, defects in certain ADC capacitors and resistors are difficult to detect a using a partial linearity test. Thus linearity tests, including partial linearity tests, are time consuming, costly, and may have trouble catching defects in certain ADC capacitors and resistors.